The Next Dimension

From Zork to Zero: The Infocom Story

Wayne Williams

It was the company responsible for bringing text adventures into the home, the company which for five years dominated the American software charts with classic games such as Zork and Hitchhiker's Guide to the Galaxy, and the company which in the face of falling sales gambled on diversifying its product range and failed spectacularly. Wayne Williams chronicles the dramatic rise and fall of Interactive Fiction legends Infocom

Dave Lebling and Marc Blank loved Colossal Cave, the famous first text adventure game written by Willie Crowther and added to four years later by Don Woods. The two MIT (Massachusetts Institute of Technology) students were seriously addicted to the game. So addicted in fact that in 1977 they decided to write their own version, convinced they could make a better, more natural world. Lebling and Blank were part of a set called the Dynamic Modelling (DM) group within MIT’s Laboratory for Computer Science, although they weren’t majoring in computers –- Lebling was doing a degree in political science, while Blank was training to become a doctor. With fellow DM student Tim Anderson they began by creating a small four-room world using a parser that Lebling had written in MDL (Muddle), a programming language created in the MIT labs as a successor for LISP.

Buoyed by the success of their first foray into adventure-game writing they roped in a fourth student, Bruce Daniels, and began work on what was eventually to become Zork. (Zork, incidentally, was a nonsense word used in the MIT labs in the 1970s, as was Frob, a word that was later used in Zork II: The Wizard of Frobozz). Although they made quick progress on designing the world and its famous story, it was to be another two years –- and after plenty of feedback from fellow students who could log in and play the game six at a time via the university’s PDP-10 mainframe –- before the game was finally finished. There was no commercial aim behind the creation of Zork; it was just a bit of fun.

Infocom is born

Al Vezza, professor at MIT, Assistant Director of LCS and leader of the DM group, had for a while been interested in putting together a company to make money from computer programming. Anderson, Lebling and Blank were all keen to continue to work with one another, so it seemed natural for them to join forces. Another professor, JCR Licklider (known as Lick), who had raised funding for the LCS projects, was also interested in being part of the new venture.

Infocom, a name chosen purely because of its inoffensiveness to everyone, was officially founded on 22nd June 1979 by 10 LCS members in total, all of whom committed money –- ranging from $400 up to $2,000 -– to buy their shares and get the company off the ground. The founding members were: Tim Anderson, Joel Berez, Marc Blank, Mike Broos, Scott Cutler, Stu Galley, Dave Lebling, JCR Licklider, Chris Reeve and Al Vezza. The original board of directors consisted of Lebling, Vezza, Broos, Berez and Galley, although in November both Lebling and Galley resigned from the board and were replaced by Blank and Licklider. Shortly after Broos, who was president, also stepped down and Berez took over. No one worked full-time for the company at this point; it was a part-time project with everyone keeping their day job or completing their studies. The company office was a PO Box.

The idea behind Infocom was to produce computer software. What kind of computer software no one knew. Ideas such as databases, office programs and medical software were all bandied about. Eventually Anderson and Lebling suggested the company retail Zork as it was already written and tested. That made a lot of sense to the rest of the board and so it was agreed that Zork would become Infocom’s first official release.

The initial problem the fledgling company faced was how to get its massive 1Mb mainframe game to fit on to the 32Kb home machines. Thankfully Berez and Blank had independently come up with the first part of the answer –- a multiformat emulator known as the Z-Machine Interpretive Program, or ZIP for short. ZIP would be different for each machine it ran on but its aim would be the same -– to run a virtual processor called Z-Machine. And Z-Machine’s role was equally straightforward. As well as compressing text (using 5.5 bits per character instead of the usual 8) it would run the Zork Implementation Language, or ZIL, an updated version of Lebling’s original Zork parser that the game would be written in. This meant each computer platform would simply need a one-off ZIP writing for it; the games wouldn’t need recoding from scratch. This was something that would come in very handy when the company later began producing a lot of games.

Of course even all this foresight and inventiveness didn’t entirely solve the problem of space -– the other answer did that. To get a 1Mb game on to a 32Kb platform, huge chunks would have to be chopped out of it. Zork promptly became a game in three parts.

Look who's Zorking

Personal Software (later known as VisiCorp), publisher of the Visicalc spreadsheet, agreed to distribute Zork I for Infocom. In November 1980 the first version of the game, for the PDP-11, hit the shelves. It was followed just one month later by the first home version, for the TRS-80. Bruce Daniels, now working for Apple, wrote a ZIP for the Apple II and that became the third format to be released.

Although the game was moderately successful, selling 1,500 copies for the TRS-80 and 6,000 copies for the Apple II, Infocom wasn’t too happy with Personal Software’s commitment or how the game had been marketed. The original box art, with a moustachioed warrior hacking and slashing away, misrepresented the nature of Zork which was all about clever thinking and puzzle solving. Infocom, which had always intended to just be a software developer not a game publisher, decided it could market its own products better and bought back the remaining $32,000’s worth of stock. It then repackaged the disks to reflect the true nature of the game and in October 1981 started selling it itself. A month later Zork I was joined on the shelves by the inevitable Zork II, and Marketing Manager Mort Rosenthal joined the company, which now had its first office in Boston, Massachusetts. Berez and Blank also became its first fulltime employees.

The two Zork games did very well, and with $160,000 in sales by the end of 1981, Infocom was on the up. Two new games followed: Blank, who loved detective novels, wrote Deadline, then came the third Zork adventure. Steve Meretzky, who would later go on to create many of Infocom’s most famous games, joined in November 1981 as a tester on Deadline after the existing one, Meretzky’s roommate Michael Dornbrook, went off temporarily to the University of Chicago’s business administration programme. Dornbrook stayed involved in Infocom however. Noticing the number of letters the company received from players begging for help, he had previously set up the Zork User Group, a $2 per-hint service and a newsletter known as the New Zork Times, and with his father’s help continued to run it out of a Milwaukee PO Box.

Dornbrook’s most famous innovation was the idea of InvisiClues –- solutions to the games’ trickiest puzzles written in invisible ink which could be revealed using a special pen. This was an idea that was suggested to Mike at a party by a friend after he bemoaned the problem of sending the same hints and solutions out time after time. Two manufacturers in the US were capable of producing the books and pens, and luckily one of them was based nearby.

InvisiClues were sold through bookshops and were a massive success, going on to make up most of the top 10 in the computer book charts –- until (following complaints from other publishers) the chart compilers lumped them together as one publication. At this point InvisiClues simply maintained a stranglehold on the number one position. Each game came with a coupon that allowed the player to buy an InvisiClue booklet and a complete map for $4.95 (at cost). If you didn’t have the coupon, perhaps because you’d pirated the software for example, it would cost you $8.95. By 1983, when the Zork User Group was finally absorbed into Infocom, it had over 20,000 members.

A fateful decision

On New Year’s Day 1982 the company moved to another office in Cambridge, Massachusetts. Encouraged as the board was by the huge success of its games, it was also jealous of the real money being made by business software developers. Infocom’s adventures sold for under $50 dollars each, while business programs went for around 10 times that amount. In particular Infocom was envious of Lotus Development which had been created by some friends from MIT (and Mitch Kapor, the man who had distributed Zork I for Infocom at Personal Software). Many people within Infocom felt the company should be looking to diversify, and so in October 1982 a new business division was created.

This decision wasn’t all that strange, given that Infocom’s original raison d’etre had been to produce general software. Some of the older members of the company, Vezza in particular, were a little uncomfortable with running a firm that only made games; they wanted to do something slightly more serious. They also felt that Infocom’s current success was a bubble that wouldn’t last forever, and that they needed to branch out in order to continue to grow. It was a commercial decision that made a lot of sense.

Brian Berkowitz and Richard Ilson, who had worked together in an LCS group called Project MAC on the same floor as the DM group at MIT, were thinking of building a relational database and took the idea to Infocom because they knew the people there. This coincided with Infocom’s wish to do something different, and so it looked like fate. The first product from this new venture was swiftly announced –- a relational database called Cornerstone. To help Infocom’s transition to a producer of serious software, Berez became Chief Operating Officer and Vezza took over as Chief Executive Officer.

In the meantime, game after game followed. All names that adventure fans will be more than familiar with: Suspended, The Witness, Planetfall (Steve Meretzky’s first game), Infidel, and Enchanter (developed with the working title Zork IV). Every game was a massive hit, selling upwards of 100,000 copies each. Adverts at the time proclaimed “Infocom –- The Next Dimension”, and the games clogged up the sales charts. By the end of 1983 all of the company’s games were in the top 40 Softsel chart, and the three-year-old Zork I was still at the very top ahead of the likes of Lode Runner, Zaxxon, Frogger, Ultima III and Microsoft Flight Simulator. By the end of 1983 Infocom’s annual sales totalled more than $6 million.

1984 saw Infocom’s most famous success. Hitchhiker’s Guide to the Galaxy creator Douglas Adams was a big fan of the company’s work, having been introduced to its adventures with Suspended. He was keen to have his books turned into a game and knew Infocom was the company to do it. Meretzky was assigned the job of working with Adams because of his experience in the sci-fi genre (Planetfall having garnered plenty of best game awards in 83/84) and the two got together for numerous discussions on the game, both face-to-face and electronically.

Hitchhiker’s was produced with the two authors working remotely -– Steve in Massachusetts, Douglas in London -– on two DEC System 20s (the computer all Infocom’s adventures were written on at the time) hooked up over the Dialcom network. The game came out in 1984 and was the company’s biggest success since Zork. A sequel was discussed, planned and even tentatively started a few years later, but the sales from Bureaucracy, Adams’ only other Infocom game, were poorer than hoped so the idea was shelved.

Running into trouble

By the end of 1984, when the firm moved to new and very expensive premises at 125 Cambridge Park Drive, Infocom’s annual sales were up to $10 million. Development on the increasingly costly Cornerstone was continuing and by now the increase in programmers, marketing and admin staff had brought Infocom’s number of employees up to 100.

Infocom sought outside Venture Capital funding, but its mix of games and planned business software found few takers. Gulf and Western, owners of Simon and Schuster, offered $20 million for the entire games side of the company, but the offer was rejected. In the end Infocom only managed to secure $500,000, but remained confident that its database could be funded by the future revenue from its games sales.

After a long time in development the $495 program finally debuted in 1985 to rave reviews. Cornerstone was ahead of its time with plenty of well-thought-out and innovative features. It had friendly menus, users could add descriptions to files and fields and it would only allow you to enter the correct information (only company names in the company field for example). It was able to autocorrect spelling errors, multivalue and variable length fields were supported, and it was compatible with all the other databases of the day, including Lotus 123, dBase II, and Symphony. In keeping with Infocom’s more famous products it also supported parsing and could recognise friendly date statements such as ‘Next Monday’. It also fitted comfortably onto a single floppy disk.

Unfortunately, for all its pluses, Cornerstone also had problems. Unlike dBase II, it wasn’t programmable -– you had to use the built-in functions. Also, by 1985 the PC was becoming the dominant computer platform. Cornerstone’s use of a virtual machine to make it platform independent (like the adventure games) meant it ran slowly, especially when handling larger databases.

The software went on to sell well, clocking up 10,000 sales in its first year. Unfortunately, the cost of producing it –- $2.5 million -– coupled with the fact that it brought in $1.8 million in sales, rather than the projected $4.7 gross profit, made it an expensive failure. This alone probably wouldn’t have killed off Infocom, but the company’s games were also not doing as well as previously. The text adventure market had begun to die. Graphical adventures, on the next generation of computers like the Amiga and Atari, were rapidly becoming the way forward. Infocom’s 1985 annual sales totalled $10 million, the same as the year before, but way below its projected revenue of $12 million plus.

Redundancies were unavoidable. However, the losses were so great the business division was forced to close down entirely. In an attempt to resurrect Cornerstone and at least get back some of the outlay, the software’s price was dropped to $99.95, with little success. It was a serious blow to Infocom and Al Vezza who had been convinced the company could “out-Lotus, Lotus”.

Activision

On 13th June 1986, Infocom, by now down to 40 employees (many of whom had taken pay cuts to keep the company afloat), was purchased by Activision for $7.5 million. At first many believed this to be a good thing. It meant the company could continue at least. Unfortunately for Infocom, six months after the takeover Activision’s CEO Jim Levy was replaced by Bruce Davis. Davis had been against the takeover from the start and set out to make life as difficult, and costly, for the new acquisition as possible. Infocom’s games now had a much shorter shelf life and the company was expected to produce eight of them a year, rather than the usual four or five, but with no extra staff. Sales fell to around 10,000 copies per game.

In 1986 Infocom bowed down to public pressure and introduced graphics into its games for the first time. Fooblitzky, a multiplayer board game in which four players (as dogs) competed against one another in a race to collect the correct four objects, was the first of these. It had graphics, but they were two-colour and poorly drawn. The game unsurprisingly flopped.

In 1987 the company released Brian Moriarty’s Beyond Zork, its first game with a graphical user interface (of sorts), including a map, a separate window for the inventory items and various RPGstyle additions. A year later Steve Meretzky’s Zork Zero came out packed with full-colour VGA delights, including graphical games and puzzles. Unfortunately, these concessions to changing public tastes weren’t enough to save the company, and in May 1989 Activision forced Infocom to lay off 15 of its remaining 26 staff and move to the company’s headquarters in Silicon Valley. With just five of the remaining Infocom staff members prepared to make the move, the truth was undeniable. Infocom, as we knew and loved it, was dead.

>Steve Meretzky interview

Infocom’s most famous Imp, now principal game designer at online-gaming portal WorldWinner.com, talked to us about his memories of the great company and what it was like working with Douglas Adams.

Retro Gamer: How did you get involved with Infocom?

Steve Meretzky: I already knew several of the founders, Marc Blank and Joel Berez. We were all involved with the film program at MIT. I was also rooming with Mike Dornbrook, who was Infocom’s first (and, at the time, only) game tester. When he left town to attend business school in Chicago, Marc asked me if I’d like to replace him as Infocom’s playtester. I tested Deadline, Zork III, and Starcross, and then Marc offered me the chance to become a game writer (‘implementor’, in Infocom parlance).

RG: What is your favourite memory of working at Infocom?

SM: There are just so many –- it was such a young, fun group of people and we were more like a big extended family than a company. A lot of my favourite memories revolve around parties. We had beer every Friday at 5pm, but often these would be expanded to fully fledged parties or some sort of unique festivity. For instance, the time we put on a mock trial of Hollywood Dave for killing all the goldfish in the fish pond (in an attempt to clean the pond). Or the time we had a graduation ceremony for our CEO, Al Vezza, after he attended a weeklong 'CEO School'. Or the Halloween party where we put on an 'interactive play', with audience members shouting parser-like instructions to the cast.

Another fun thing we often did at these Friday parties was to hold Infocom versions of TV game shows, such as the 20,000 Zorkmid Pyramid. And for a period of several months, the Friday parties included hermit crab races, complete with parimutuel betting [where the gambler bets against other gamblers, not the house].

There were also incredible parties that the company would throw for the press at trade shows, such as the Bring Your Own Brain party at the Field Museum in Chicago, where we sent invitees on a scavenger hunt through the museum for prizes. Or another party at the Field Museum, where we hired the Second City cast to create skits promoting Stationfall and The Lurking Horror. Or the party at Elvis Presley’s former mansion in Las Vegas, with the Murder-To-Go mystery troupe staging a killing to promote our latest game, Suspect.

Another great memory is the pre-Christmas season when our orders were running far ahead of the production facility’s ability to assemble the game boxes. So employees signed up for Saturday and Sunday shifts to keep the assembly lines running, operating labelling machines, shrinkwrapping machines, etc. It was a great example of the whole company pulling together but also having a fun time.

But certainly the best memory I have is meeting the lovely Betty Rock, who worked in the sales department at Infocom and who will join me in celebrating our 19th anniversary in a few days!

RG: Congratulations! Tell us about Planetfall. What was it like designing your first game?

SM: Slow at first, as I had to learn a new programming language (ZIL). Also, at the beginning, I was still spending about half my time testing. So even though I started writing the game in September of 1982, it didn’t really start coming together until perhaps early spring of ‘83. But I quickly learned that writing games was even more fun than testing them.

RG: Were you surprised by people’s emotional response to Floyd’s death?

SM: I wasn’t surprised that people had an emotional response; I wrote that scene to create one. But I was certainly surprised by the magnitude of the response. And by the fact that people are still asking about it, 20 years later!

RG: What was the game design process like in general?

SM: It varied a lot from game to game. If a game was a sequel –- such as my second game, Sorcerer –- it was necessary to follow the content and style of the previous game or games. In that case, where Sorcerer was the fifth game in the sequence, after the Zork Trilogy and Enchanter, there was a huge amount of world background and history to be consistent with, so I spent a long time combing through the code for those four games, compiling a detailed database of Zork lore. (I continued to maintain that database through subsequent games, and ended up including it within a game –- Zork Zero in 1988 -– as referenced through a copy of the Encyclopedia Frobozzica.)

If a game was an adaptation of material from another medium, as in the case of The Hitchhiker’s Guide to the Galaxy, that also brought a series of constraints on the design process.

Sometimes one element of the design -– the storyline, the geography, the puzzles, the humour, the package elements –- would be dominant and would be developed first, with the other elements then formed around that. But other times they would all be fairly equal and would evolve in parallel.

RG: How did you come to work with Douglas Adams?

SM: Douglas was an Infocom player and fan, and so when he, his agent and his publisher began discussing the subject of a computer-game adaptation of Hitchhiker’s Guide, he was adamant that it be with Infocom. Marc Blank suggested that I collaborate on the game with Douglas, partly due to fortunate timing (I had just completed Sorcerer), partly because many people had found Planetfall to be reminiscent of the humour of Hitchhiker’s Guide, and partly because I was the only implementor who was as tall as Douglas.

RG: What was he like?

SM: The best way to describe Douglas is that he was the ideal dinner companion. He could speak intelligently and with wit about almost any topic under the sun. Unfortunately, as a collaborator, he suffered from the fact that he was the world’s worst procrastinator! I had to practically camp out on his doorstep in England to get him to finish his stuff for the game. Otherwise, working with him was great. He had such a different perspective on things, and came up with puzzles and scenes that I’d never have thought of in a million years on my own -– having the game lie to you, or using a parser failure as the words which fell through a wormhole in the universe and started an interstellar war, or having an object like "no tea".

RG: How did the infamous Babel fish puzzle originate?

SM: The basic idea was by Douglas, and I added some refinements (like the Upper-Half-Of-The-Room Cleaning Robot). More interesting is how close the puzzle came to being removed from the game; most of Infocom’s testing group thought it was too hard. I was going into a meeting with them just as Douglas was leaving for the airport at the end of his final trip to Infocom, and I asked him, "What should I tell them about the Babel fish puzzle?" He said, "What should you tell them? Tell them to fuck off!" So the puzzle stayed... and its very hardness became a cult thing. Infocom even sold T-shirts that said “I got the Babel fish!"

RG: What are your memories of working on the Hitchhiker’s game?

SM: Around May of 1984, with the game just a few weeks away from its deadline for the start of alpha testing, and about half the game still undesigned, I went over to England. Douglas was not only procrastinating on the game, he was also procrastinating on the fourth Hitchhiker’s book, So Long and Thanks For All the Fish. His agent had sent him to a country inn in western England, far from the distractions of London life. That’s where I went, with instructions to camp out on his doorstep until the game design was done. We spent four days at this really pleasant inn, a former baronial mansion, sipping expensive wines and designing the game. How can life get any better than that?

RG: What is your favourite Infocom game?

SM: My favourite Infocom game has always been Dave Lebling’s Starcross, but I also love Jeff O’Neill’s Nord and Bert Couldn’t Make Head or Tail of It for being so original and different. Of my own games, my favourite would be A Mind Forever Voyaging, because it was my largest, most serious, and most socially relevant work, and because I feel it showed that computer games could be more than an adolescent pastime, but could instead be used to explore Big Issues. Except perhaps for Floyd’s death, it’s the game of mine that seems to have touched people most deeply.

RG: Were there any games that never got finished? (You were working on a Titanic game that got cancelled we believe?)

SM: Not many during Infocom days, but many since then. I never really started the Titanic game at Infocom, although it probably would have been my next game after Zork Zero if Activision hadn’t shut Infocom down at that point. I tried selling the Titanic idea numerous times during post-Infocom days, in particular when I had my own development studio (Boffo Games), but was told over and over that "no one's interested in the Titanic". So it was with a bit of irony that I watched Cameron’s movie become the highest grossing film in history.

RG: Have you ever had the inclination to write another text adventure? (Such as Leather Goddesses 3 or Planetfall 2: The Search for Floyd?)

SM: Actually, I wrote a design for a new Planetfall game for Activision back in 1993 and 1994. It asked for rewrite after rewrite, and finally just killed the game. Every couple of years, I hear a rumour that they’re going to revive it. But I would like to write a text adventure at some point when I have some time... Perhaps after the kids go to college...

---

---

These historical, out-of-print articles and literary works have been GNUSTOed onto InvisiClues.org for academic and research purposes.